Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 12288
    5. 8-RES_FS12288-JB
    6. 979
    Prev
    Next
    INDEX
    Explanations

    phrases related to perspectives or opinions

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    97
    +0.10
    0.4%
    289
    +0.09
    0.4%
    705
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    705
    +0.10
    0.03
    576
    +0.09
    0.02
    341
    +0.09
    0.02
    Negative Logits
    usters
    -0.79
    ongo
    -0.76
    oute
    -0.70
     livest
    -0.70
    ividual
    -0.69
    uster
    -0.67
    ynski
    -0.67
    erville
    -0.66
    iry
    -0.66
    unts
    -0.66
    POSITIVE LOGITS
     Sabha
    0.81
    fare
    0.78
    ward
    0.76
    forward
    0.74
    finding
    0.70
    ï
    0.69
     forever
    0.67
     Alma
    0.62
    footed
    0.62
     Shank
    0.62
    Activations Density 0.064%

    No Known Activations